Mitral Leaflet Changes Following Myocardial Infarction: Clinical Evidence for Maladaptive Valvular Remodeling

Abstract

Background

Ischemic MR is classically ascribed to functional restriction of normal leaflets, but recent studies have suggested post-MI MV leaflet fibrosis and thickening, challenging valve normality. Progression of leaflet thickness post-MI has not been studied. We hypothesized that excessive MV remodeling post-MI contributes to MR. Our objectives are to characterize mitral valve (MV) changes following myocardial infarction (MI) and relate them to mitral regurgitation (MR).

Methods and Results

Three groups of 40 patients with serial echocardiograms over a mean of 23.4 months were identified from an echocardiography database: patients first studied early (6±12 days) and late (12±7 years) after an inferior MI, and normal controls. MV thickness was correlated with MR. We studied the mechanisms for MV changes in a sheep model (6 apical MI vs 6 controls) followed 8 weeks, with MV cellular and histopathological analyses. Early post-MI, leaflet thickness was similar to controls (2.6±0.5vs2.5±0.4 mm, p=0.23), but significantly increased over time (2.5±0.4 to 2.9±0.4 mm, p<0.01). In this group, patients tolerating maximal doses of renin-angiotensin blocking agents had less thickening (25% of patients, p<0.01). The late-MI group had increased thickness (3.2±0.5vs2.5±0.4 mm, p<0.01) without progression. At follow-up, 48% of post-MI patients had more than mild MR. Increased thickness was an independently associated with MR. Experimentally, 8 weeks post-MI, MVs were 2-fold thicker than controls, with increased collagen, pro-fibrotic TGF-β, and endothelial-to-mesenchymal transformation, confirmed by flow cytometry.

Conclusion

MV thickness increases post-MI and correlates with MR, suggesting an organic component to ischemic MR. MV fibrotic remodeling can indicate directions for future therapy.

Ischemic mitral regurgitation (MR) is a frequent¹ complication of myocardial infarction (MI), doubling heart failure and mortality^2–4. Its primary mechanism is leaflet tethering by disturbed left ventricle (LV) geometry secondary to local or global remodeling^5–16. Ischemic MR is considered “functional”, without contribution from intrinsic leaflet changes. The size of the valve was classically assumed to be fixed in the remodeling ventricle; however, the mitral valve (MV) can enlarge to adapt to ventricular expansion^17–21 and reduce MR¹⁸. Experimental studies have demonstrated that mechanical stretch induced by leaflet tethering can induce active valve growth^{19, 22}. This compensatory mechanism, however, is frequently unable to compensate for LV distortion, and ischemic MR remains prevalent^{18, 21}.

Other work suggested that abnormal leaflets could contribute to functional MR: increased valve stiffness and collagen accumulation were observed in end-stage heart failure, suggesting organic MV alteration ^{20, 23–25}. Based on collagen synthesis upregulation in response to stress²⁵, Kunzelman et al. showed by finite-element analysis that increased leaflet stiffness and thickness can interfere with MV function²⁶. Normal closure requires flexible leaflets to bend and form a coaptational seal, and normal MV can expand up to 15% in systole to close (Figure 1)^{27, 28}; these elements of normal function could be limited with stiffer leaflets^{26, 29}. Mechanisms and clinical implications of these fibrotic changes have not been investigated. Especially, it is unclear if and how initially compensatory stretch-induced valve growth can progress towards maladaptive stiffening and thickening contributing to MR. In the setting of MI, numerous processes such as renin-angiotensin system (RAS) activation are known to affect LV remodeling; but whether or not leaflet tissue can be involved in the remodeling process is unknown ^30–33.

Figure 1.

Normal mitral valve closure requires flexible leaflets to form a coaptational seal.

We tested the hypothesis that post-MI MV changes could result in maladaptive fibrosis with excessive thickening contributing to MR. The prevalence and timing of valve thickening and their relation to MR were assessed. We studied over time three groups of patients: those with recent acute MI, those with remote MI and normal controls. In a subsidiary study, we used a sheep model of MI to explore cellular and molecular mechanisms leading to fibrosis. In order to capture changes due to MI without the papillary muscle displacement seen with infero-posterior MI or the associated MR, which can secondarily thicken the leaflets^{34, 35}, we created limited antero-apical MIs that do not generate MR³³.

Methods

1. Human retrospective follow-up study (Figure 2)

Figure 2.

Clinical study design. Three groups were followed for serial echocardiographic characterization of mitral leaflets.

We used our institution’s echocardiography database to identify patients with documented MI involving at least the inferior or posterior walls, leaflet tethering and at least 2 echocardiograms separated by 90 days or more ⁵. The first 40 consecutive patients were studied in each of three groups: 1) Patients with an echo within the first three weeks of a first MI and a follow-up study at least 90 days later (early-MI group); 2) Patients with remote (>5 years) MI and comparable follow-up (late-MI group); and 3) Control subjects (comparable age and sex to both MI groups) with similar follow-up to control for changes in leaflet thickness with aging ³⁶. The control group consisted of patients with serial normal echoes either after foramen ovale closure or screening for chemotherapy-induced cardiomyopathy. Exclusion criteria were >mild aortic stenosis or insufficiency, MV organic pathology (prolapse, rheumatic, endocarditis or extensive annular calcification), valve prosthesis, Marfan disease and patient taking anti-parkinsonian dopamine agonists.

Echocardiographic analysis (Figure 3)

Figure 3.

Representative examples showing mitral leaflet evolution after MI, thickness and leaflet excursion measurements. A-B: Immediately after MI, parasternal and apical views showing a thin valve with full diastolic excursion. C-D: Same patient 5 years later. The mitral leaflets are thicker, with limited excursion. The systolic function and ventricular dimensions did not change between the two exams. E: Another patient with initially thin and mobile valve (video 1) with definite thickening, reduced excursion and severe MR 3 years later (F and G, video 2 and 3). Displayed frames show the maximal diastolic leaflet opening.

All echocardiograms were reviewed blinded to the patient group and timing of the exam. MV thickness was measured in parasternal long-axis and apical zoomed views in a diastolic frame without valve motion, with leaflets perpendicular to the echocardiographic beam, taking advantage of the axial resolution. This was achieved at mid-diastole in parasternal view and end-diastole in apical views. Thickness was measured three times for both leaflets in both views in areas free of chordal attachment; all values were averaged^37–39. A subset of 10 patients was measured by 2 investigators, and 1 month later by the first investigator for inter- and intraobserver variabilities. MV excursion angle between end-systole and full opening was measured in the parasternal long-axis view⁴⁰. LV end-diastolic and end-systolic dimensions (LVEDD and LVESD) were measured and LVEF calculated (modified Quinones method). MR was graded independently by two level 3 readers (discrepancies were resolved by a third reader as needed) as suggested by current guidelines with an integrated approach using all available parameters^{41, 42}. MR grade was defined as 1 (no or trace), 2 (mild: effective regurgitant orifice [ERO]<20 mm², regurgitant volume [RV] <30 ml, vena contracta [VC]<0.3 cm), 3 (moderate: ERO 0.20-0.39 mm², RV 30–59 ml, VC 0.3-0.69 cm) or 4 (severe: ERO>40 mm², RV≥60 ml, VC≥0.7 cm). In this population, more than mild MR (as defined by ERO>20 mm² or RV>30 ml) has been shown to have a prognostic impact³. For that reason, greater than mild MR was considered significant in our study. Progression of MR was defined as an increase of at least one grade of MR between baseline and follow-up studies. Medical records were reviewed for clinical characteristics and medication profile. The study was approved by the institutional review committee.

2. Experimental study

Six adult Dorsett hybrid sheep underwent left anterior descending artery ligation, and six had sham thoracotomy. Epicardial echocardiography was done at baseline and sacrifice. MVs were harvested at 8 weeks for analyses. Thickness (averaged over 10 thickest mid-leaflet sites), valve morphology and collagen accumulation were analyzed by Hematoxylin/Eosin and Masson’s trichrome staining. Cellular activation (endothelial-to-mesenchymal transformation [EMT]) was assessed by staining for endothelial (CD31) and interstitial myofibroblast (α-smooth muscle actin [α-SMA]) markers. EMT was confirmed by flow cytometry of dissociated fresh valve endothelial cells with fluorescent anti-CD31 and anti-α-SMA labeling. Transforming growth factor (TGF-β) was assessed by immunohistochemistry as an initial measure of growth signaling promoting both EMT and fibrosis.

Statistics

Continuous variable were expressed as mean±standard deviation, and categorical variables as number (%). Differences in means were tested for significance with Student’s T-tests and differences in proportions with Chi-square tests. Differences in thickness over time among groups were assessed with repeated-measures ANOVA and paired T-test. Variables associated with >mild MR for all post-MI patients at follow-up were assessed by logistic regression. Leaflet thickness, LVEF, LVEDD, LVESD, age, leaflet excursion, time from infarct date, tethering distances (from papillary muscles to annulus), left atrial dimension and annulus dimensions (apical two- and four-chamber views) in mid-systole¹⁴ were tested in univariate analysis, and variables with p<0.05 were tested in multivariable model. Inter- and intra-observer agreements of leaflet thickness measurements were assessed using a single measure, two-way random effect intraclass correlation coefficient (ICC). Values of 0.893 (interobserver) and 0.946 (intraobserver) were obtained. Correlation and Bland-Altman plots are presented in Supplementary Figure 1. In the animal model, EMT by flow cytometry and echocardiographic measures were compared between MI and control groups with t-tests. Statistical analysis was performed with Stata/IC 11.2 (StataCorp LP, Tx).

RESULTS

1. Human studies

A total of 120 patients were studied, each with two echocardiographic studies. The first echocardiogram was performed 6±12 days post-MI (85% within one week) in the early-MI group and 12±7 years post-MI in the late-MI group. There was no significant difference in age, gender or median follow-up time (642, 544 and 773 days) among groups. Both MI groups had lower LVEF and larger LV dimensions than normal (Table 1). Most patients had revascularization at the time of MI (early-MI: 33/40; late-MI: 36/40). Primary PCI was the preferred revascularization strategy for both groups (70% and 60% for early-MI and late-MI groups). Reasons for non-revascularization included late presentation without ongoing chest pain, absence of reversible ischemia on non-invasive testing or patients refusing invasive procedures.

Table 1.

Control, Late MI and Early MI groups baseline characteristics

	Control (n=40)	Early-MI (n=40)	Late-MI (n=40)	p value
Time post-MI	-	6±12 days	12±7 years	-
Age	66±13	66±11	70±10	0.17
Male, n(%)	26(65)	25(63)	30(75)	0.45
LVEF(%)	69±7	38±10*	41±11*	<0.01
LVEDD(mm)	46±6	49±6	55±7*†	<0.01
LVESD(mm)	29±5	37±8*	44±7*†	<0.01
Tethering Distance, PPM (mm)	38±5	41±5	42±5	<0.01
Tethering distance, LPM (mm)	37±5	40±5	43±6	<0.01
MR>mild	0(0)	13(33)*	19(48)*	<0.01
Anterior thickness(mm)	2.7±0.6	2.7±0.5	3.4±0.6*†	<0.01
Posterior thickness(mm)	2.5±0.5	2.3±0.5	3.0±0.5*†	<0.01
Average thickness(mm)	2.6±0.5	2.5±0.4	3.2±0.6*†	<0.01
Excursion anterior leaflet(degree)	64±14	41±15*	38±14*	<0.01
Excursion posterior leaflet(degree)	62±19	47±17*	38±14*	<0.01
Hypertension	23(58%)	27(68%)	31(78%)	0.16
Diabetes	2(5%)	10(25%)*	14(35%)*	<0.01
Hyperlipidemia	19(48%)	25(63%)*	35(88%)*	<0.01
Renal insufficiency	2(5%)	6(15%)*	15(38%)*†	<0.01
ACEI/ARB	20(50%)	37(93%)*	30(75%)*	<0.01
Beta-blockers	15(38%)	38(98%)*	36(90%)*	<0.01
Aspirin	24(60%)	37(93%)*	34(85%)*	<0.01
Statin	19(48%)	37(93%)*	36(90%)*	<0.01

LVEF: Left ventricle ejection fraction; LVEDD: left ventricle end-diastolic diameter; LVESD: left ventricle end-systolic diameter; PPM: posterior papillary muscle; LPM: lateral papillary muscle; MR: mitral regurgitation; ACEI: angiotensin converting enzyme inhibitors; ARB: angiotensin receptor blockers.

^*p<0.05 vs control;

^†p<0.05 vs early-MI.

Mitral valve thickness and motion (Figure 4)

Figure 4.

Leaflet thickness at two time points in control, early-MI and late-MI groups. Early-MI group shows significant progression over time. Late-MI group presents increased but stable thickness. *: p<0.01 vs control and Early-MI (baseline and follow-up); †: p<0.01 vs Early-MI Baseline.

Average MV thickness was stable over time in the control group (2.6±0.5 mm to 2.6±0.5 mm, p=0.71). The early-MI group had initially similar thickness to controls (2.5±0.4 vs 2.6±0.5 mm, p=0.23), but showed significant progression over time (2.5±0.4 mm to 2.9±0.4mm, p<0.001). The proportion of early-MI patients with thickness >3 mm increased from 13% at baseline to 43% at follow-up (p<0.01, see example in Figure 3 and supplementary video 1–3). Late-MI patients had thicker leaflets compared to early-MI and control patients at baseline (3.2±0.5 vs 2.5±0.4 vs 2.6±0.5 mm for late-MI, early-MI and controls, p<0.01). The proportion of patients with increased thickness was higher in the late-MI group (68%, 13% and 18% of patients had thickness >3 mm in late-MI, early-MI and control groups, p<0.01). Despite increased thickness at baseline, the late-MI group had no progression over time. Opening excursion for both leaflets was stable over time in the control group. At baseline, the early-MI group had decreased excursion compared with controls, and there was an additional significant decrease at follow-up (Table 2, Figure 3). The late-MI group had reduced leaflet excursion at baseline compared to control and early-MI groups, but was stable over time. There was no significant difference in leaflet thickness for patients with vs without revascularization or post-MI ischemia by non-invasive tests (p=NS for all). There was no observed difference in thickness for patients with associated comorbidities (diabetes, renal failure, hypertension, hyperlipemia or active smoking). No difference in thickness was observed based on the echocardiography system used (51% Philips ie-33; 31% Philips Sonos 7500; 18% GE Vivid 7).

Table 2.

Early-MI population at baseline and follow-up

	Baseline	Follow-up	p value
Days post-MI	6±12	773±558
LVEF(%)	38±10	38±13	0.83
LVEDD(mm)	49±6	53±7	<0.01
LVESD(mm)	37±8	42±8	<0.01
Tethering Distance, PPM (mm)	41±5	42±5	0.13
Tethering distance, LPM (mm)	40±5	41±4	0.23
Annulus AP4(mm)	35±4	36±3	<0.01
Annulus AP2(mm)	35±4	35±3	0.6
MR>2	13(33%)	19(48%)	0.25
Anterior thickness(mm)	2.7±0.5	3.1±0.5	<0.01
Posterior thickness(mm)	2.3±0.5	2.6±0.4	<0.01
Average thickness(mm)	2.5±0.4	2.9±0.4	<0.01
Excursion anterior leaflet(degree)	41±15	37±17	0.03
Excursion posterior leaflet(degree)	47±17	40±19	0.01

LVEF: Left ventricle ejection fraction; LVEDD: left ventricle end-diastolic diameter; LVESD: left ventricle end-systolic diameter; PPM: posterior papillary muscle; LPM: lateral papillary muscle; MR: mitral regurgitation; AP4: apical 4-chamber; AP2: apical 2-chamber.

Effect of medication on thickness (Figure 5)

Figure 5.

Top: thickness progression early post-MI in patients taking high vs low ACEI/ARB doses. Bottom: post-MI MV thickness according to the presence of MR.

Medication profile is shown in Table 1: 75% of late-MI and 93% of early-MI patients were treated with angiotensin converting enzyme inhibitors (ACEI) or angiotensin receptor blockers (ARB). However, the tolerated dose was low (less than 50% of maximal dose for a given medication) for a significant proportion of patients (66/80). In an exploratory analysis, we tested the hypothesis that leaflet thickness progression could be different in patients taking higher doses of ACEI or ARB. In the subgroup of early-MI patients taking higher doses, there was no significant leaflet thickening (high doses: 2.6±0.4 mm to 2.7±0.6 mm, p=0.44; low doses: 2.5±0.4 mm to 2.9±0.4 mm, p<0.01). In the late-MI group, anterior leaflets were thinner in the high-dose subgroup (anterior thickness: 2.9±0.6 vs 3.4±0.5 mm, p=0.04) However, our study was not designed and had limited power to assess this difference.

Association of leaflet thickness and ischemic MR (Figure 5)

In the pooled ensemble of post-MI patients, those with increase of at least one grade of MR at follow-up vs baseline had also significant increase in average leaflet thickness (+0.43±0.46 vs +0.06±0.46 mm, for patients with vs without MR progression, p=0.002). At follow-up, there was a significant association between leaflet thickness and greater than mild MR (average thickness 3.2±0.5 vs 2.8±0.4, p=0.0006 for patients with versus without MR), without difference in LVEF (p=0.36) or LV dimensions (p=0.48 for LVEDD and 0.99 for LVESD, Supplementary Table). Age (p=0.003) and left atrial dimension (p=0.001) were significantly associated with MR in univariate analysis. In multivariate analysis, average thickness (β coefficient±SE: 1.47±0.62, p=0.018) and age (β coefficient±SE: 0.058±0.027, p=0.027) were the factors associated with greater than mild MR, while LA dimension was not (p=0.85).

2. Experimental mechanistic study

All animals survived until sacrifice; none developed MR. There was a mild decrease in LVEF post-MI (61±7 to 47±4%, p<0.01). MV thickness (microscopy) was significantly increased post-MI vs sham (1.02±0.23 vs 0.44±0.11 mm, p<0.01). Histopathology post-MI showed expansion of the central spongiosa layer and focal subendothelial deposition of collagen, primarily on the atrial leaflet surface (Figure 6).

Figure 6.

Mitral valves thickness (microscopy) in sham (A) and post-MI (B) sheep. Post-MI valves (C) are significantly thicker (p<0.01).

Cellular changes (Figure 7)

Figure 7.

Left: Molecular histopathology showing positive staining for CD31 and α-SMA at and beneath the atrial endothelial border consistent with endothelial-mesenchymal transformation post-MI. Top right: Flow cytometry showing increase in endothelial cells co-expressing α-SMA in post-MI valves vs sham. Bottom right: Strong staining for TGF-β in post-MI valve.

The endothelial layer of control MVs had CD31+ cells without α-SMA staining. In contrast, post-MI MVs were positive for both CD31 and α-SMA, indicating EMT. By flow cytometry, endothelial cells coexpressing α-SMA were more common in post-MI vs sham MVs (48±14% versus 7±4% of endothelial cells, p<0.01). The endothelium and subendothelial interstitium were strongly positive for TGF-β1, colocalized to regions of α-SMA staining.

Discussion

In this study, we demonstrate that MV presents echocardiographic changes post-MI: 1) Early post-MI, leaflet thickness is initially normal, but increases over time; 2) Late post-MI, thickness is maximal without progression. Although aging is associated with valve thickening³⁶, lack of increase in a control group of comparable age, sex and follow-up time indicates this is not a likely explanation for the changes observed early post-MI.

The capacity for MV remodeling in functional MR has been demonstrated previously^{17–19, 23, 24, 26, 27, 35}. While active valve expansion can be seen as adaptive, excessive remodeling can lead to fibrosis with increased thickness, decreased mobility, and potentially more MR since effective closure requires systolic expansion and flexibility²⁶. Although differences between the late-MI and control groups could be related to other comorbidities, the evolution of changes in the early-MI cohort and the large-animal study are consistent with changes beginning only after MI, mainly in the early period, as the late-MI group showed stable thickness.

Mechanistic considerations

Ischemic MR is the complex result of mechanical stretch in an ischemic environment with subsequent heart failure-related humoral activation. A previous study from our group demonstrated that mechanical stretch alone from papillary muscle displacement causes active valve enlargement¹⁹. In the experimental portion of the current study, we show that an apical MI without papillary muscle involvement also triggers MV alterations. Importantly, strong TGF-β staining post-MI was not observed in the stretch-only model, suggesting a specific role for the ischemic environment on MV remodeling. While MR itself could contribute to the observed changes through increased turbulence and shear stress^{34, 35}, the animal cohort showed a cluster of histologic changes in the absence of MR.

Activation of RAS is well known post-MI⁴³, and angiotensin II can trigger TGF-β expression in the heart^44–46, with subsequent collagen deposition and extracellular matrix remodeling. Our observations suggest that valve leaflets can be involved in the global post-MI remodeling; and increased TGF-β likely plays a role in the observed leaflet changes (Figure 8). Although our clinical study was not primarily designed to assess the effect of medication on MV thickness, the observed relation between ACEI/ARB doses and leaflet thickness post-MI deserves attention as it is consistent with TGF-β involvement: RAS blockade is known to inhibit pro-fibrotic effects of TGF-β in various organs^{47, 48} including blood vessels and myocardium^{49, 50}. Our data suggest a possible effect of TGF-β on MV remodeling, which could represent a pharmacologic target. Although we cannot directly link MR to these histological changes in this experimental work, these observations warrant further investigations as this could be a first step towards medical approaches targeting leaflet fibrosis to prevent ischemic MR.

Figure 8.

Proposed mechanisms for post-MI adverse leaflet remodeling. Endothelial-to-mesenchymal transformation creates more interstitial cells, and TGF-β triggers collagen deposition.

Clinical significance of increased leaflet thickness

This study challenges the concept that ischemic MR is exclusively related to LV remodeling. It is reasonable to suggest that disturbed extracellular matrix can change the mechanical properties of the valve, increasing its stiffness^{23, 26}. This can limit MV closure by decreasing systolic expansion, limiting flexible leaflet bending, and potentially decreasing adaptive valve growth as compensatory enlargement is attenuated in these patients^{18, 21}. All MI patients had the substrates for ischemic MR: decreased EF, infero-posterior MI and MV tethering. However, not all of them had significant MR, suggesting other variables in the pathogenesis of MR. Leaflet thickness was a significantly associated with MR, and in the evolving early-MI group, increased leaflet thickness was associated with MR progression.

Limitations

The human data are retrospective, and unknown factors not accounted for could have explained differences in thickness between groups, but not likely within the early-MI group over time. MR quantification can change according to loading conditions. Our population was limited to pre-selected patients (inferior MI and leaflet tethering) more likely to have ischemic MR. Difference in individual follow-up timing, absence of 3D valve metrics and relatively small sample size are other limitations. Non-invasive ways to measure fibrosis and biomechanical properties of cardiac valves are limited, but a finite-element model previously demonstrated that increased thickness can impair coaptation²⁶. Although leaflet thickness was assessed non-invasively in the clinical study, our protocol (averaging 12 measures/patient) showed good reproducibility and correlated well with previous studies³⁶; leaflet thickening post-MI was confirmed by pathology in the sheep study. Despite small sample size in the animal study, observed differences were consistent and highly significant. Our animal protocol included apical MI without MR (contrasting with the clinical study: inferior MI with high ischemic MR prevalence); this was important to demonstrate that the observed leaflet changes are not limited to inferior MI or MR itself. Interestingly, decreased diastolic excursion paralleled the increase in thickness post-MI. Although this can be attributed to decreased cardiac output or diastolic tethering⁴⁰, intrinsic leaflet changes could also explain this phenomenon in part and could be explored in future biomechanical studies.

Conclusion

MV undergo multiple changes post-MI. Excessive valve remodeling can result in maladaptive fibrosis suggesting an organic component in ischemic MR. The role of TGF-β and RAS to modulate this remodeling merits exploration.

Clinical Perspective.

Ischemic mitral regurgitation is generally considered “functional”: normal leaflets unable to close properly in a dilated and distorted left ventricle. However, recent experimental work in animal models suggested leaflet abnormalities potentially contributing to functional mitral regurgitation. In this manuscript, we compare the progression of mitral valve thickness in three groups of patients: those with 1) recent and 2) remote inferior myocardial infarction, and 3) normal controls. We show that progressive mitral valve thickening occurs early after myocardial infarction, and is associated with later mitral regurgitation. In an associated animal experiment, we demonstrate mitral valve thickening following even a limited apical myocardial infarction, with evidence of fibrotic remodeling and strong presence of TGF-β in the leaflets. This clinical and experimental work suggests an organic component to ischemic mitral regurgitation and suggests mitral leaflet remodeling as a potential therapeutic target.

Abbreviations list

MR

Mitral Regurgitation

MI

Myocardial Infarction

LV

Left Ventricle

MV

Mitral Valve

LVEF

Left Ventricle Ejection Fraction

EMT

Endothelial-to-Mesenchymal Transformation

α-SMA

Alpha-Smooth Muscle Actin

TGF-β

Transforming Growth Factor-β

ACEI

Angiotensin Converting Enzyme Inhibitor

ARB

Angiotensin Receptor Blockers